close Warning: Can't synchronize with repository "(default)" (/project/cmucl/public_html/git/cmucl.git does not appear to be a Git repository.). Look in the Trac log for more information.

Changes between Version 1 and Version 2 of Release20f


Ignore:
Timestamp:
10/04/14 21:18:33 (3 years ago)
Author:
rtoy
Comment:

Fix up markup.

Legend:

Unmodified
Added
Removed
Modified
  • Release20f

    v1 v2  
    2626    * Accuracy of trig functions improved. (Previously, they could be
    2727      off by one or two ulp.)
    28     * RUN-PROGRAM accepts :ELEMENT-TYPE to allow specifying the
    29       element type for streams that are created for RUN-PROGRAM.  (See
     28    * {{{RUN-PROGRAM}}} accepts {{{:ELEMENT-TYPE}}} to allow specifying the
     29      element type for streams that are created for {{{RUN-PROGRAM}}}.  (See
    3030      ticket:92)
    3131
    3232  * Changes
    3333    * Update to ASDF 3.1.3.
    34     * When *PRINT-CASE* is :DOWNCASE, integers are printed with
     34    * When {{{*PRINT-CASE*}}} is {{{:DOWNCASE}}}, integers are printed with
    3535      lowercase letters when needed.
    36     * Micro-optimize KERNEL:DOUBLE-FLOAT-BITS for x86/sse2.
    37     * Add micro-optimization for unary FTRUNCATE for x86/sse2.
    38     * Compiler can derive the types of the results of DECODE-FLOAT.
    39     * Main disassembly routines in the DISASSEMBLE package accept
    40       :BASE, :CASE, and :RADIX keyword arguments, like
    41       DISASSEMBLE:DISASSEMBLE.
     36    * Micro-optimize {{{KERNEL:DOUBLE-FLOAT-BITS}}} for x86/sse2.
     37    * Add micro-optimization for unary {{{FTRUNCATE}}} for x86/sse2.
     38    * Compiler can derive the types of the results of {{{DECODE-FLOAT}}}.
     39    * Main disassembly routines in the {{{DISASSEMBLE}}} package accept
     40      {{{:BASE}}}, {{{:CASE}}}, and {{{:RADIX}}} keyword arguments, like
     41      {{{DISASSEMBLE:DISASSEMBLE}}}.
    4242    * Micro optimizations for sse2 to allow descriptor regs and stack
    4343      tns for many operations so that the operands can be in memory
     
    4545      instruction.
    4646    * Support for darwin/ppc 10.5 and gcc 4.8.
    47     * Micro-optimization:  (expt -1 power) just returns -1 or 1
     47    * Micro-optimization:  {{{(expt -1 power)}}} just returns -1 or 1
    4848      without actually computing the power.
    4949    * Run-time checks for zero pages on newly allocated pages has been
     
    5858      to fdlibm instead of implementing them in Lisp.
    5959    * Source location information has been added for
    60       DEFINE-CONDITION. (From Helmut Eller.)
     60      {{{DEFINE-CONDITION}}}. (From Helmut Eller.)
    6161    * The lisp executable is now compiled to use SSE2 on x86 machines;
    6262      CMUCL will not run on chips without SSE2 anymore.
    63     * (cosh 1000d0) signals an overflow error as it
     63    * {{{(cosh 1000d0)}}} signals an overflow error as it
    6464      should. Previously, it just incorrectly returned infinity.
    65     * Deprecating FLOAT-TRAPPING-NAN-P in favor of
    66       FLOAT-SIGNALING-NAN-P.  These kinds of numbers are called
     65    * Deprecating {{{FLOAT-TRAPPING-NAN-P}}} in favor of
     66      {{{FLOAT-SIGNALING-NAN-P}}}.  These kinds of numbers are called
    6767      signaling NaNs, not trapping NaNs.
    6868    * Changed the printer to print "Signaling" instead of "Trapping"
     
    7878      failing if cmucl is unable to interpret the instruction that
    7979      caused the exception.
    80     * Micro-optimization for sparc in rounding a double-float to a
     80    * Micro-optimization for sparc in rounding a {{{DOUBLE-FLOAT}}} to a
    8181      32-bit signed integer.
    82     * The compiler can now derive the type of ROUND. (But more work is
     82    * The compiler can now derive the type of {{{ROUND}}}. (But more work is
    8383      needed.)
    84     * Fix an issue with multi-processing destroy-process doesn't take
     84    * Fix an issue with multi-processing {{{DESTROY-PROCESS}}} doesn't take
    8585      effect for a schedulable process because it has no run-reason
    8686      and/or arrest-reason.
     
    9090      hyperbolic functions have been updated.  Instead of following
    9191      the text in the CLHS, the values are defined according to the
    92       definitions in the formulas.  In particular (atanh x) for x real
     92      definitions in the formulas.  In particular {{{(atanh x)}}} for x real
    9393      and on the branch cut is now continuous with different
    9494      quadrants.  This differs from the description of the branch cut
    95       for atanh in the CLHS.
    96     * CLEAR-OUTPUT was not actually doing anything. Now, CLEAR-OUTPUT
     95      for {{{atanh}}} in the CLHS.
     96    * {{{CLEAR-OUTPUT}}} was not actually doing anything. Now, {{{CLEAR-OUTPUT}}}
    9797      will discard any buffered data that has not been written out
    9898      yet.
     
    108108    * The float constants for various limits are defined for
    109109      double-double floats such as
    110       least-positive-double-double-float.
     110      {{{LEAST-POSITIVE-DOUBLE-DOUBLE-FLOAT}}}.
    111111    * When reading very small float values, round to
    112112      least-positive-foo-float if possible.
    113     * (log -0w0) and (log 0w0) returns values analogous to the
     113    * {{{(log -0w0)}}} and {{{(log 0w0)}}} returns values analogous to the
    114114      double-float versions.
    115     * Fix bug in printing MOST-NEGATIVE-FIXNUM.
    116     * For Gray streams, CLEAR-OUTPUT was calling STREAM-FORCE-OUTPUT
    117       instead of STREAM-CLEAR-OUTPUT, so the output wasn't actually
     115    * Fix bug in printing {{{MOST-NEGATIVE-FIXNUM}}}.
     116    * For Gray streams, {{{CLEAR-OUTPUT}}} was calling {{{STREAM-FORCE-OUTPUT}}}
     117      instead of {{{STREAM-CLEAR-OUTPUT}}}, so the output wasn't actually
    118118      cleared.
    119     * In some situations KERNEL:DOUBLE-FLOAT-BITS on x86 would cause a
     119    * In some situations {{{KERNEL:DOUBLE-FLOAT-BITS}}} on x86 would cause a
    120120      segfault. This has been fixed.
    121121    * For Linux, motifd is no longer a 64-bit app.
    122     * (exp 1d0) now returns the correctly rounded value of
     122    * {{{(exp 1d0)}}} now returns the correctly rounded value of
    123123      e. Previously, it was off by one bit.
    124     * INTEGER-DECODE-FLOAT returns the correct values for denormal
    125       doubles. As a side-effect of this fix, DECODE-FLOAT returns the
    126       correct values for denormals, and SCALE-FLOAT scales denormals
     124    * {{{INTEGER-DECODE-FLOAT}}} returns the correct values for denormal
     125      doubles. As a side-effect of this fix, {{{DECODE-FLOAT}}} returns the
     126      correct values for denormals, and {{{SCALE-FLOAT}}} scales denormals
    127127      correctly.
    128     * EXT:UNIX-NAMESTRING no longer returns NIL if a directory does
     128    * {{{EXT:UNIX-NAMESTRING}}} no longer returns {{{NIL}}} if a directory does
    129129      not exist. This was a regression from at least 18a.
    130130